PROJECT SUMMARY DNA sequencing and new computational approaches have yielded detailed maps of clonal variation in human cancer. While changes in clonal structure over time and under the selective pressure of treatment have been extensively studied in hematologic malignancies, solid cancers are less well characterized owing to the relative lack of suitable tumor material. Analyses of breast and ovarian cancer have demonstrated substantial clonal variation between metastatic sites and polyclonal heterogeneity within individual tumor deposits, yet our understanding of the dynamics of clonal change in breast and ovarian cancer and its role in therapeutic response and the emergence of resistance is in its infancy. By combining expertise in mutation detection and genomic analysis with access to unique patient cohorts, this proposal will develop critically needed methods to identify all genomic changes in tumors in order to resolve a tumor's clonal substructure as it evolves over time or space in response to treatment. We will apply our tools in two key patient cohorts: 1) longitudinal samples from early stage, neoadjuvant breast cancer patients biopsied before, during, and after the completion of initial chemotherapy; and 2) tumor cells from metastatic breast and ovarian cancer patients at multiple time-points during their treatment with multiple courses of chemotherapy (breast and ovarian) and at time of autopsy (ovarian). The Specific Aims are to: (1) Develop and apply comprehensive mutation detection to identify the genetic lesions that develop in patient tumors over time during the course of chemotherapy, or at multiple distinct metastatic lesions. Using these tools, we will measure the cellular prevalence of mutations among multiple biopsies from both breast and ovarian patient cohorts. (2) Comprehensively prioritize mutations based on the likelihood that they drive tumor evolution. We will use these methods to prioritize consequential mutations and to gain insight into the potential mechanisms underlying clonal evolution. (3) Delineate tumor subclone structure and its evolution across longitudinal tumor biopsies and multiple metastatic lesions. By estimating the cellular prevalence of all forms of mutation in each biopsy, these innovations will enable a better understanding of how tumor subclone populations evolve over time and space and evade response to chemotherapy. (4) Create an interactive, web-based software platform for the analysis exploration of tumor subclone structure. In summary, the proposed research will devise and apply new algorithms that will improve our understanding of the dynamics of breast and ovarian cancer evolution over time and space.